This package implements the [HSU], [HSD], [AHSU], [AHSD] and [HBR-lambda] procedures for discrete tests (see References).
The functions are reorganized from the reference paper in the following way.
discrete.BH (for Discrete Benjamini-Hochberg) implements
[HSU], [HSD], [AHSU] and [AHSD] and DBR (for Discrete
Blanchard-Roquain) implements [HBR-lambda]. DBH and ADBH
are wrappers for discrete.BH to access [HSU] and [HSD], as well as
[AHSU] and [AHSD] directly. Their main arguments are a vector
of raw observed p-values, and a list
of the same length, which elements are the discrete supports
of the CDFs of the p-values.
The function fisher.pvalues.support allows to compute
such p-values and support in the framework of a Fisher's
exact test of association. It has been inspired by an help
page of the package discreteMTP.
The function fast.Discrete is a wrapper for fisher.pvalues.support
and discrete.BH which allows to apply discrete procedures
directly to a data set of contingency tables.
We also provide the amnesia data set, used in
our examples and in our paper. It is basically the amnesia data set
of package discreteMTP, but slightly reformatted (the difference lies in column 3).
No other function of the package should be used, they are only internal functions called by the main ones.
D<U+00F6>hler, S., Durand, G., & Roquain, E. (2018). New FDR bounds for discrete and heterogeneous tests. Electronic Journal of Statistics, 12(1), 1867-1900. doi:10.1214/18-EJS1441.